TravelEval - Fictitious Attraction Rate: leaderboard
Metric: Fictitious Attraction Rate (FAR): recommended attractions that do not exist in the sandbox (%); Direct prompting (one pass, no tools), plans simulated in the TravelEval sandbox of 10 Chinese cities (real rail, flight, hotel and attraction data, queuing-time model, road-network distances) over 1,150 queries; lower is better. Source: arxiv.org. Saturation forecast: Estimated already saturated. 7 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 3 8B | 2.49 |
| 2 | Claude Sonnet 4.5 | 2.92 |
| 3 | Gemini 2.0 Flash | 3.64 |
| 4 | GPT-4o (2024-11-20) | 3.67 |
| 5 | GPT-4o Mini (2024-07-18) | 3.9 |
| 6 | DeepSeek V3.1 (Non-reasoning) | 4.37 |
| 7 | GPT-5 Chat | 4.51 |
Interactive version: theaggregate.ai/benchmark?slug=traveleval-fictitious-attraction-rate · How It Works · Data refreshed daily, snapshot 2026-09-29.